Propranolol (Del Frari 2018) • nlmixr2lib

Model and source

Citation: Del Frari L, Leaute-Labreze C, Guibaud L, Barbarot S, Lacour J-P, Chaumont C, Delarue A, Voisard J-J, Brunner V. Propranolol pharmacokinetics in infants treated for Infantile Hemangiomas requiring systemic therapy: Modeling and dosing regimen recommendations. Pharmacol Res Perspect. 2018;6(3):e00399. doi:10.1002/prp2.399.
Description: One-compartment population PK model for oral propranolol in infants (aged 50-243 days, 3.6-9.7 kg) with proliferative Infantile Hemangiomas (Del Frari 2018). First-order absorption and first-order elimination; apparent oral clearance CL/F scales with body weight using a fixed allometric exponent of 0.75 and a reference weight of 6.3 kg (the median weight pooled across visits D1-D84). Apparent volume V/F has no covariate effect (the paper tested but did not retain weight on V/F). Between-subject variability is retained on CL/F and Ka only; BSV on V/F was dropped from the final model (large 95% CI including 0 and 62.3% eta-shrinkage). Residual error is proportional.
Article: https://doi.org/10.1002/prp2.399

Population

Del Frari 2018 developed a one-compartment population PK model for oral propranolol in 22 infants with proliferative Infantile Hemangiomas recruited from four French hospitals (Bordeaux, Lyon, Nantes, Nice; EudraCT 2009-018102-22). Infants were stratified into two groups by age at inclusion: Group 1 (n = 10) was 35-90 days old at inclusion with the steady-state PK assessment performed after 4 weeks of treatment (D28); Group 2 (n = 12) was 91-150 days old at inclusion with the steady-state PK assessment performed after 12 weeks of treatment (D84). 167 plasma concentrations contributed to the final analysis: 21 predose troughs at D7 and 18 at D14 during titration, then 60 intensive samples in Group 1 at D28 and 68 in Group 2 at D84 (predose, 1, 2, 4, 6, and 9 h post-dose).

The pooled cohort had median age 104 days (range 50-151) and median weight 6.3 kg across all visits D1-D84. Mean weight at the PK assessment was 5.62 kg in Group 1 (D28) and 7.78 kg in Group 2 (D84). Treatment was a pediatric oral solution of propranolol hydrochloride titrated over 2 weeks (1 mg/kg/day, then 2 mg/kg/day) to the target 3 mg/kg/day BID dose for weeks 3-12. Sample baseline demographics are in Del Frari 2018 Table 2.

The same information is available programmatically: readModelDb("DelFrari_2018_propranolol")$population.

Source trace

Per-parameter origin (also recorded as in-file comments next to each ini() entry of inst/modeldb/specificDrugs/DelFrari_2018_propranolol.R):

Equation / parameter	Value	Source location
`lcl`	`log(19.3)`	Del Frari 2018 Table 5: CL/F = 19.3 L/h at WT = 6.3 kg (95% CI 15.4-23.2, RSE 10.3%)
`lvc`	`log(122)`	Del Frari 2018 Table 5: V/F = 122 L (95% CI 104-140, RSE 7.32%); no covariate retained
`lka`	`log(0.993)`	Del Frari 2018 Table 5: Ka = 0.993 1/h (95% CI 0.558-1.43, RSE 22.4%)
`e_wt_cl`	`fixed(0.75)`	Del Frari 2018 Results 3.4: allometric exponent on CL/F fixed to 0.75 on AIC grounds; reference WT = 6.3 kg per Equation 5
`etalcl`	`0.195`	Del Frari 2018 Table 5: omega^2 on CL = 0.195 (44.2% CV; 95% CI 0.0492-0.341, RSE 38.2%)
`etalka`	`1.75`	Del Frari 2018 Table 5: omega^2 on Ka = 1.75 (132% CV; 95% CI 0.370-3.13, RSE 40.2%)
`propSd`	`sqrt(0.0953) = 0.309`	Del Frari 2018 Table 5: proportional sigma^2 = 0.0953 (30.9% CV; 95% CI 0.0675-0.123, RSE 14.9%)
`(WT / 6.3)^0.75` on CL	n/a	Del Frari 2018 Equation 5; reference weight 6.3 kg = pooled median across D1-D84 (paragraph below Eq 5 in Section 3.4)
`d/dt(depot)`, `d/dt(central)`	n/a	Del Frari 2018 Results 3.3: best structural model is a one-compartment disposition with first-order absorption and first-order elimination
`Cc <- central / vc * 1000`	n/a	Dose mg / volume L -> mg/L; ng/mL = mg/L * 1000 to match the paper’s reporting units
`Cc ~ prop(propSd)`	n/a	Del Frari 2018 Results 3.3: residual variability described by a proportional error model

Virtual cohort

Original observed concentrations are not publicly available. The simulation below replicates the Del Frari 2018 study cohort: 10 Group 1 infants at the D28 median weight (5.6 kg) and 12 Group 2 infants at the D84 median weight (7.65 kg). Each infant is replicated 50 times by Monte Carlo to give 1100 simulated trajectories per dosing regimen (the source paper replicated the n = 22 cohort 1000 times to derive Table 6 percentiles; 50 replicates keeps the vignette inside the pkgdown 5-minute per-vignette wall-time budget while preserving the qualitative Cmin / Cmax features).

set.seed(20260529L)

n_replicates <- 50L
tau          <- 12      # h; primary BID dosing interval
n_days_ss    <- 7L      # 7 days to reach steady state (~12 elimination half-lives)
total_dur    <- n_days_ss * 24

# Per-subject weights for the published cohort. Group 1 median weight at
# D28 (5.6 kg) and Group 2 median weight at D84 (7.65 kg) per Del Frari
# 2018 Table 2 D28 and D84 rows.
subject_specs <- dplyr::bind_rows(
  tibble::tibble(group = "Group 1", WT = 5.6,  n = 10L),
  tibble::tibble(group = "Group 2", WT = 7.65, n = 12L)
)

# Build the per-replicate subject set with disjoint IDs.
build_subjects <- function() {
  ids <- integer(0)
  wt  <- numeric(0)
  grp <- character(0)
  off <- 0L
  for (i in seq_len(nrow(subject_specs))) {
    n_i <- subject_specs$n[i]
    ids <- c(ids, off + seq_len(n_i))
    wt  <- c(wt,  rep(subject_specs$WT[i], n_i))
    grp <- c(grp, rep(subject_specs$group[i], n_i))
    off <- off + n_i
  }
  tibble::tibble(id = ids, WT = wt, group = grp)
}

# Build dosing + observation events for a single regimen.
# `dose_times_h` is the vector of clock-times-since-midnight at which doses
# are taken on a single day; this is repeated for `n_days_ss` days, with
# dose amount = total_daily_mgkg / length(dose_times_h) * WT per dose.
make_cohort <- function(regimen_label, dose_times_h,
                        total_daily_mgkg = 3, id_offset = 0L) {
  subj <- build_subjects() |>
    dplyr::mutate(rep_id = id_offset + id) |>
    dplyr::select(id = rep_id, WT, group)

  per_dose_mgkg <- total_daily_mgkg / length(dose_times_h)
  dose_offsets  <- tidyr::expand_grid(
    day  = 0:(n_days_ss - 1L),
    hour = dose_times_h
  ) |>
    dplyr::mutate(time = day * 24 + hour) |>
    dplyr::pull(time)

  dose_rows <- tidyr::expand_grid(
    id   = subj$id,
    time = dose_offsets
  ) |>
    dplyr::left_join(subj, by = "id") |>
    dplyr::mutate(
      amt  = per_dose_mgkg * WT,
      evid = 1L,
      cmt  = 1L
    )

  obs_grid <- seq(0, total_dur, by = 0.5)
  obs_rows <- tidyr::expand_grid(
    id   = subj$id,
    time = obs_grid
  ) |>
    dplyr::left_join(subj, by = "id") |>
    dplyr::mutate(
      amt  = 0,
      evid = 0L,
      cmt  = NA_integer_
    )

  dplyr::bind_rows(dose_rows, obs_rows) |>
    dplyr::mutate(regimen = regimen_label) |>
    dplyr::arrange(id, time, dplyr::desc(evid))
}

regimens <- list(
  list(label = "Regular BID (8h, 20h)",   doses = c(8, 20)),
  list(label = "Irregular BID (8h, 17h)", doses = c(8, 17)),
  list(label = "TID (8h, 14h, 20h)",      doses = c(8, 14, 20)),
  list(label = "TID (8h, 12h, 20h)",      doses = c(8, 12, 20))
)

# Build all 50 replicates of each regimen with disjoint IDs across the
# whole simulation.
n_subj_per_rep <- sum(subject_specs$n)
events <- do.call(
  dplyr::bind_rows,
  lapply(seq_along(regimens), function(r_idx) {
    reg <- regimens[[r_idx]]
    do.call(
      dplyr::bind_rows,
      lapply(seq_len(n_replicates), function(rep_idx) {
        off <- ((r_idx - 1L) * n_replicates + (rep_idx - 1L)) * n_subj_per_rep
        make_cohort(
          regimen_label    = reg$label,
          dose_times_h     = reg$doses,
          total_daily_mgkg = 3,
          id_offset        = off
        )
      })
    )
  })
)

stopifnot(!anyDuplicated(unique(events[, c("id", "time", "evid")])))

Simulation

mod <- rxode2::rxode2(readModelDb("DelFrari_2018_propranolol"))
#> ℹ parameter labels from comments will be replaced by 'label()'

sim <- rxode2::rxSolve(
  mod,
  events = events,
  keep   = c("WT", "group", "regimen")
) |>
  as.data.frame()

Typical-clearance check (Section 3.4 anchor values)

Del Frari 2018 Section 3.4 reports the typical CL/F for the median infant in each group: 17.2 L/h at 5.6 kg (Group 1, D28) and 22.3 L/h at 7.65 kg (Group 2, D84). The packaged model uses the published CL/F = 19.3 L/h at the pooled reference weight 6.3 kg with a fixed allometric exponent of 0.75 on body weight. The table below reproduces the per-group typical clearance directly from the model’s lcl and e_wt_cl parameters.

ini_vals <- mod$ini

cl_typical <- function(wt, lcl_val, e_wt_val, ref_wt = 6.3) {
  exp(lcl_val) * (wt / ref_wt) ^ e_wt_val
}

lcl_value  <- ini_vals$est[ini_vals$name == "lcl"]
ewt_value  <- ini_vals$est[ini_vals$name == "e_wt_cl"]

typical_cl_tbl <- tibble::tribble(
  ~group,     ~wt_kg, ~published_cl_Lh,
  "Group 1",     5.6,             17.2,
  "Group 2",    7.65,             22.3
) |>
  dplyr::mutate(
    sim_cl_Lh = round(cl_typical(wt_kg, lcl_value, ewt_value), 2),
    pct_diff  = round(100 * (sim_cl_Lh - published_cl_Lh) / published_cl_Lh, 1)
  )

knitr::kable(
  typical_cl_tbl,
  caption = paste(
    "Typical apparent oral clearance CL/F (L/h) at the per-group median",
    "weight, compared to the values reported in Del Frari 2018 Section 3.4."
  )
)

Typical apparent oral clearance CL/F (L/h) at the per-group median weight, compared to the values reported in Del Frari 2018 Section 3.4.
group	wt_kg	published_cl_Lh	sim_cl_Lh	pct_diff
Group 1	5.60	17.2	17.67	2.7
Group 2	7.65	22.3	22.33	0.1

Replicate Figure 7: simulated steady-state profiles under 4 dosing regimens

Del Frari 2018 Figure 7 plots the 5th, 50th, and 95th percentile envelopes of simulated propranolol concentration over a 24-hour steady-state window for four dosing regimens at a total daily dose of 3 mg/kg/day: (A) regular BID 8:00 + 20:00, (B) irregular BID 8:00 + 17:00, (C) TID 8:00 + 14:00 + 20:00, and (D) TID 8:00 + 12:00 + 20:00. The figure below reproduces those envelopes from the packaged model.

last_day_start <- (n_days_ss - 1L) * 24
sim_ss <- sim |>
  dplyr::filter(time >= last_day_start, !is.na(Cc)) |>
  dplyr::mutate(time_of_day = time - last_day_start)

ss_pi <- sim_ss |>
  dplyr::group_by(regimen, time_of_day) |>
  dplyr::summarise(
    Q05 = stats::quantile(Cc, 0.05, na.rm = TRUE),
    Q50 = stats::quantile(Cc, 0.50, na.rm = TRUE),
    Q95 = stats::quantile(Cc, 0.95, na.rm = TRUE),
    .groups = "drop"
  )

ggplot(ss_pi, aes(time_of_day, Q50)) +
  geom_ribbon(aes(ymin = Q05, ymax = Q95), alpha = 0.25, fill = "steelblue") +
  geom_line(linewidth = 0.6, colour = "steelblue4") +
  facet_wrap(~ regimen) +
  labs(
    x = "Time of day (h since 00:00 on day 7)",
    y = "Propranolol plasma conc. (ng/mL)",
    title = "Replicates Del Frari 2018 Figure 7",
    caption = paste(
      "Median + 5-95% PI at steady state on day 7,",
      "1100 simulated trajectories per regimen (n = 22 cohort x 50 replicates)."
    )
  )

Comparison against Table 6 simulated Cmin / Cmax percentiles

Del Frari 2018 Table 6 reports the 5th / 50th / 95th percentiles of simulated Cmin and Cmax over a 24-hour steady-state window for the four dosing regimens above. The table below reproduces those percentiles from the packaged model and renders them alongside the published anchors. Per Del Frari 2018 Section 3.6, simulated median Cmax differed by less than 10% across regimens and median Cmin by less than 20%; the packaged-model simulations should reproduce that qualitative finding.

cmin_cmax_sim <- sim_ss |>
  dplyr::group_by(regimen, id) |>
  dplyr::summarise(
    Cmin = min(Cc, na.rm = TRUE),
    Cmax = max(Cc, na.rm = TRUE),
    .groups = "drop"
  ) |>
  dplyr::group_by(regimen) |>
  dplyr::summarise(
    sim_cmin_p05 = round(stats::quantile(Cmin, 0.05, na.rm = TRUE), 2),
    sim_cmin_p50 = round(stats::quantile(Cmin, 0.50, na.rm = TRUE), 2),
    sim_cmin_p95 = round(stats::quantile(Cmin, 0.95, na.rm = TRUE), 2),
    sim_cmax_p05 = round(stats::quantile(Cmax, 0.05, na.rm = TRUE), 2),
    sim_cmax_p50 = round(stats::quantile(Cmax, 0.50, na.rm = TRUE), 2),
    sim_cmax_p95 = round(stats::quantile(Cmax, 0.95, na.rm = TRUE), 2),
    .groups = "drop"
  )

published_t6 <- tibble::tribble(
  ~regimen,                       ~pub_cmin_p05, ~pub_cmin_p50, ~pub_cmin_p95, ~pub_cmax_p05, ~pub_cmax_p50, ~pub_cmax_p95,
  "Regular BID (8h, 20h)",                 1.99,          14.0,          39.4,          24.0,          61.8,           131,
  "Irregular BID (8h, 17h)",              0.803,          11.6,          47.0,          26.4,          67.1,           141,
  "TID (8h, 14h, 20h)",                    1.43,          14.2,          50.2,          24.7,          62.2,           130,
  "TID (8h, 12h, 20h)",                    1.53,          15.2,          43.4,          23.6,          60.1,           128
)

t6_compare <- dplyr::left_join(cmin_cmax_sim, published_t6, by = "regimen")

knitr::kable(
  t6_compare,
  caption = paste(
    "Simulated steady-state Cmin and Cmax percentiles (5 / 50 / 95)",
    "by dosing regimen, alongside Del Frari 2018 Table 6 published anchors.",
    "All concentrations in ng/mL."
  )
)

Simulated steady-state Cmin and Cmax percentiles (5 / 50 / 95) by dosing regimen, alongside Del Frari 2018 Table 6 published anchors. All concentrations in ng/mL.
regimen	sim_cmin_p05	sim_cmin_p50	sim_cmin_p95	sim_cmax_p05	sim_cmax_p50	sim_cmax_p95	pub_cmin_p05	pub_cmin_p50	pub_cmin_p95	pub_cmax_p05	pub_cmax_p50	pub_cmax_p95
Irregular BID (8h, 17h)	0.93	12.27	55.03	40.03	75.37	124.49	0.803	11.6	47.0	26.4	67.1	141
Regular BID (8h, 20h)	2.30	17.48	58.44	35.83	69.10	114.15	1.990	14.0	39.4	24.0	61.8	131
TID (8h, 12h, 20h)	2.06	15.79	53.89	36.30	67.77	110.65	1.530	15.2	43.4	23.6	60.1	128
TID (8h, 14h, 20h)	1.75	16.97	62.22	35.06	65.66	115.38	1.430	14.2	50.2	24.7	62.2	130

PKNCA validation

Steady-state NCA on the regular-BID regimen, computing Cmax, Tmax, Cmin (trough), and AUC over the 12-hour dosing interval at the end of day 7.

ss_bid <- sim |>
  dplyr::filter(regimen == "Regular BID (8h, 20h)", time >= (n_days_ss - 1L) * 24 + 8, time <= (n_days_ss - 1L) * 24 + 20, !is.na(Cc)) |>
  dplyr::mutate(
    tad       = time - ((n_days_ss - 1L) * 24 + 8),
    treatment = group
  )

ss_dose <- ss_bid |>
  dplyr::group_by(id, treatment, WT) |>
  dplyr::summarise(.groups = "drop") |>
  dplyr::mutate(
    time = 0,
    amt  = 1.5 * WT     # 1.5 mg/kg per BID dose
  )

conc_obj <- PKNCA::PKNCAconc(
  ss_bid |> dplyr::transmute(id, time = tad, Cc, treatment),
  Cc ~ time | treatment + id
)

dose_obj <- PKNCA::PKNCAdose(
  ss_dose |> dplyr::transmute(id, time, amt, treatment),
  amt ~ time | treatment + id,
  route = "extravascular"
)

intervals <- data.frame(
  start    = 0,
  end      = 12,
  cmax     = TRUE,
  cmin     = TRUE,
  tmax     = TRUE,
  auclast  = TRUE,
  half.life = TRUE
)

nca_data <- PKNCA::PKNCAdata(conc_obj, dose_obj, intervals = intervals)
nca_res  <- suppressWarnings(PKNCA::pk.nca(nca_data))

nca_summary <- as.data.frame(nca_res$result) |>
  dplyr::filter(PPTESTCD %in% c("cmax", "cmin", "tmax", "auclast", "half.life")) |>
  dplyr::group_by(treatment, PPTESTCD) |>
  dplyr::summarise(
    median = round(stats::median(PPORRES, na.rm = TRUE), 2),
    p05    = round(stats::quantile(PPORRES, 0.05, na.rm = TRUE), 2),
    p95    = round(stats::quantile(PPORRES, 0.95, na.rm = TRUE), 2),
    .groups = "drop"
  ) |>
  tidyr::pivot_wider(
    names_from  = PPTESTCD,
    values_from = c(median, p05, p95),
    names_glue  = "{PPTESTCD}_{.value}"
  ) |>
  dplyr::arrange(treatment)

knitr::kable(
  nca_summary,
  caption = paste(
    "Steady-state NCA over the 12-hour BID dosing interval on day 7 by",
    "study group (median; 5-95% PI). Cmax / Cmin in ng/mL,",
    "Tmax / half.life in h, auclast in ng/mL*h."
  )
)

Steady-state NCA over the 12-hour BID dosing interval on day 7 by study group (median; 5-95% PI). Cmax / Cmin in ng/mL, Tmax / half.life in h, auclast in ng/mL*h.
treatment	auclast_median	cmax_median	cmin_median	half.life_median	tmax_median	auclast_p05	cmax_p05	cmin_p05	half.life_p05	tmax_p05	auclast_p95	cmax_p95	cmin_p95	half.life_p95	tmax_p95
Group 1	460.03	63.90	19.47	5.22	2	217.31	33.46	2.87	2.36	0.5	1011.33	102.19	61.32	13.31	4.5
Group 2	503.47	74.47	16.49	4.22	2	250.14	39.42	2.11	2.00	0.5	972.42	116.53	54.59	9.84	4.0

Comparison against Table 3 observed steady-state concentrations

Del Frari 2018 Table 3 reports the observed median (range) propranolol plasma concentration at predose and 1, 2, 4, 6, and 9 h after the morning dose at D28 (Group 1) and D84 (Group 2). The table below extracts the corresponding times from the simulated steady-state Day 7 profile and compares to the published observed-data medians.

key_times <- c(0, 1, 2, 4, 6, 9)
sim_t3 <- sim_ss |>
  dplyr::filter(
    regimen == "Regular BID (8h, 20h)",
    time_of_day >= 8,
    (time_of_day - 8) %in% key_times
  ) |>
  dplyr::mutate(time_after_dose = time_of_day - 8) |>
  dplyr::group_by(group, time_after_dose) |>
  dplyr::summarise(
    sim_median = round(stats::median(Cc, na.rm = TRUE), 2),
    .groups = "drop"
  )

published_t3 <- tibble::tribble(
  ~group,    ~time_after_dose, ~pub_median,
  "Group 1",                0,        17.5,
  "Group 1",                1,        62.2,
  "Group 1",                2,        61.8,
  "Group 1",                4,        47.0,
  "Group 1",                6,        37.2,
  "Group 1",                9,        33.8,
  "Group 2",                0,        9.74,
  "Group 2",                1,        41.9,
  "Group 2",                2,        74.3,
  "Group 2",                4,        44.2,
  "Group 2",                6,        31.4,
  "Group 2",                9,        21.7
)

t3_compare <- dplyr::left_join(sim_t3, published_t3,
                               by = c("group", "time_after_dose")) |>
  dplyr::mutate(
    pct_diff = round(100 * (sim_median - pub_median) / pub_median, 1)
  )

knitr::kable(
  t3_compare,
  caption = paste(
    "Simulated steady-state median propranolol concentration at the Del Frari",
    "2018 Table 3 sampling times, compared against the published observed",
    "medians from Group 1 (D28) and Group 2 (D84). All values in ng/mL."
  )
)

Simulated steady-state median propranolol concentration at the Del Frari 2018 Table 3 sampling times, compared against the published observed medians from Group 1 (D28) and Group 2 (D84). All values in ng/mL.
group	time_after_dose	sim_median	pub_median	pct_diff
Group 1	0	19.47	17.50	11.3
Group 1	1	57.96	62.20	-6.8
Group 1	2	58.16	61.80	-5.9
Group 1	4	49.40	47.00	5.1
Group 1	6	39.44	37.20	6.0
Group 1	9	28.50	33.80	-15.7
Group 2	0	16.49	9.74	69.3
Group 2	1	67.48	41.90	61.1
Group 2	2	67.35	74.30	-9.4
Group 2	4	55.29	44.20	25.1
Group 2	6	42.58	31.40	35.6
Group 2	9	26.87	21.70	23.8

Assumptions and deviations

Cohort weight stratification at the per-group D28 / D84 median. The packaged validation vignette simulates 10 Group 1 infants at the D28 median weight (5.6 kg) and 12 Group 2 infants at the D84 median weight (7.65 kg), matching the sample sizes and per-group steady- state assessment timing of the source clinical study (Del Frari 2018 Table 2 D28 and D84 rows). Within-group weight variability is held at the median rather than sampled from the observed range to keep the residual variability dominated by the packaged-model between-subject IIV on CL/F and Ka, which is what the Figure 7 and Table 6 percentile envelopes reflect.
Allometric reference weight 6.3 kg. Del Frari 2018 Equation 5 and the paragraph immediately below it define the allometric reference as 6.3 kg (the median weight in the pooled population from D1 to D84). The same value is hard-coded in the model file’s (WT / 6.3)^e_wt_cl expression.
Allometric exponent fixed at 0.75. Del Frari 2018 Results Section 3.4 paragraph 2 reports that the fixed-0.75 exponent yielded a lower AIC than the estimated-exponent variant; the packaged model encodes the exponent as fixed(0.75) to preserve this status. A downstream user re-fitting the model can release this constraint but the packaged simulation should not.
No BSV on V/F. Del Frari 2018 Results Section 3.4 paragraph 4 removed BSV on V/F from the final model because the 95% CI of omega^2 included 0 and the eta-shrinkage was 62.3%. The packaged model has no etalvc term, consistent with the final Table 5 parameterisation; V/F is a fixed-effect-only parameter.
No covariate effect on V/F. Del Frari 2018 Section 3.4 paragraph 3 tested but did not retain a weight effect on V/F. The packaged model has no (WT / 6.3)^e_wt_vc scaling on the volume of distribution.
Cohort N for Figure 7 reproduction. Del Frari 2018 used 1000 replicates of the n = 22 cohort (~22,000 trajectories per regimen); the packaged vignette uses 50 replicates (1100 trajectories per regimen) to fit inside the pkgdown 5-minute per-vignette wall-time budget. The smaller N adds some Monte Carlo noise to the percentile bands but preserves the qualitative findings (Cmin / Cmax differ by less than 20% / 10% across regimens; the BID-regular and TID-regular envelopes overlap substantially).
Day 7 used as the steady-state day. The source paper simulated one week of repeated dosing at the target 3 mg/kg/day dose to reach steady state before recording Cmin / Cmax; with a terminal-phase half-life of ~5-6 h (V/F = 122 L, CL/F = 19.3 L/h at 6.3 kg yields t1/2 = ln(2) * V / CL = 4.4 h; longer with allometric scaling at lower body weights), 7 days = 14 doses is well past 5 half-lives for any plausible WT.
Residual error reported on the variance scale. Del Frari 2018 Table 5 reports the proportional residual variance sigma^2 = 0.0953 per the NONMEM $SIGMA convention; the packaged model uses propSd = sqrt(0.0953) = 0.309 to obtain the equivalent linear-scale proportional SD (~30.9% CV).
Bioavailability not identifiable. Del Frari 2018 Methods Section 2.3.1 paragraph 1 notes that absolute bioavailability could not be determined from oral-only data; the structural parameters reported in Table 5 are therefore apparent (CL/F, V/F) and the packaged model has no explicit F1 parameter. Downstream users performing dose-to-exposure simulations should treat the dose variable as an oral mg amount; the apparent-volume scaling absorbs the missing F term.